Deep Generative Networks

This is a deep learning elective focused on the generative side of vision: how neural networks can synthesize, edit, and translate images rather than just classify them. You'll spend the semester reading recent papers on GANs, diffusion-style models, and transformer-based generators, working through a homework and a midterm, and building toward a substantial project you present at the end. It assumes you already know CNNs and basic deep learning (CS 464 / 478 territory), and it's where you go if you want to do research or work in computer vision, image synthesis, or any of the modern multi-modal generative systems.